Some techniques have been developed for storing data (file) as a virtual tape volume in a hard disk drive (hereinafter referred to as “HDD”) that is a random access device, on the assumption of data storage into a sequential access device, such as a tape drive. For example, one system virtualizes a random accessible file system (e.g., a disk cache, an HDD) to cause a host to recognize the system as a tape drive that performs sequential access. Moreover, one virtual tape volume is managed as one file on a GPFS (General Parallel File System). Accordingly, in principle, such a system uses the virtual tape volume file in a sequential manner.
One technique maintains data multiplicity by building a Grid distributed system (hereinafter abbreviated as “Grid”) to copy (replicate) the virtual tape volume file among multiple clusters (nodes). Moreover, a copy function itself can be implemented to be able to accommodate a distributed environment. Specifically, each cluster requiring copy (copy target cluster) dynamically selects an optimal copy source cluster at an optimal timing, and then an actual copy is executed.
In this copy of the data in the virtual tape volume among the clusters, if the virtual tape volume has been changed (updated), it is conceivable to copy only a changed portion (difference) in order to reduce an amount of data flowing on a network among the cluster.
For example, in one system, positions of Truncate and Erase recently applied on a source cluster are recorded in something called “RST table,” and compared to the RST table as previously replicated on a target cluster, and thereby a data difference to be copied (a most recent position of data) is detected.
In another system, in a storage apparatus, a master disk controller overwrites update data stored in a second area of a cache memory so as to be transferred to a sub disk controller, with update data of the same records, for a prescribed time between first and second preset base points, and the sub disk controller handles the update data between the first and second base points as consistent actual data.